Testing out vSAN in lab.
Ran into issue while using Microsoft SQLIO tool to generate I/O load. The VM became unresponsive and inaccessible, The load test wrote 1 100GB file, on a 200GB disk, it shouldn't of 'ran out of disk'. VM is not view-able in the vSAN datastore, the datastore has "locked-in" the disk allocated to the VM, cannot recover the VM.
ESXi, vSphere versions at 5.5 U1
VM - MS Windows 2012 R2
Disk - 200GB
Test tool: MS SQLIO
Issue - Upon generating IO load using the utility - the VM became unresponsive and "inaccessible" and the vSAN storage is now "locked", not sure how to recover the storage?
esxcli - shows the volume, however its "inaccessible"
vsan 409.5G 272.0G 137.5G 66% /vmfs/volumes/vsanDatastore
/dev/disks # ls /vmfs/volumes/vsanDatastore/ -l
ls: /vmfs/volumes/vsanDatastore/a3ad0054-f20c-0a14-271e-0010183f81fc: No space left on device
lrwxr-xr-x 1 root root 36 Sep 11 14:53 VM with vSan -> a3ad0054-f20c-0a14-271e-0010183f81fc
I can't seem to recover the VM, or re-claim the storage allocated to it. Any advice how to do either, more importantly re-claim the disk.
1. What was FTT policy for VSAN? There is a good chance that it created a replica vm and left you with just enough space to crash....and you didnt specify your vsan build detsils as far as number of esxi hosts, if your on HCL, etc...in other words is there a chance it could have been the result of an unsupported build?
2. If you reboot everything or somehow get your VSAN funtioning enough to mansge, you coukd in theory add storage to one of the disk groups in vsan so that your vm would actually function. But its hard to even sepeculate eithout above details.